Nagoya 1. Med. Sci. 47. 23 - 34, 1984

STUDY OF STROMAIN SCIRRHOUS GASTRIC CARCINOMA

HIDEO KAMEl, TADASHI WATANABE,KEISUKE TERABE, TAKASHI KOJIMA

and TATSUHEI KONDO

Second Department of Surgery, Nagoya University School of Medicine,Nagoya 466, Japan

ABSTRACT

Rabbits were immunized with the supernatant of a homogenate of non-cancerous gastric mucosato obtain an antiserum. This antiserum, after absorption by human jejunal mucosa and liver tissue,produced precipitin lines with the supernatants of homogenates of gastric carcinoma and gastric mu­cosa. Fluorescent staining was attempted using this antiserum. Well-stained cytoplasms were ob­served in cells of tissue imprints of normal gastric mucosa and gastric carcinoma. When frozen sec­tions of peroperatively extirpated materials were likewise stained, the cells and the lumen of theglands were stained to a large extent in the non-cancerous gastric mucosa and tubulary adeno­carcinoma of the non-scirrhous type. In scirrhous gastric carcinoma fluorescence-stained substanceswere scantly in the lumen of the glands but observable to some degree in the tumor stroma. It issuggested that infiltration of the interstitium by glandular antigens induces proliferation of stroma ingastric carcinoma.

Keywords: Scirrhous carcinoma, Tumor stroma

INTRODUCTION

There is a variety of proliferating modes of gastric carcinoma. 1,2, 3,4) Among these, the General

Rules for the Gastric Cancer Study in Surgery and Pathology prepared by the Japanese Research

Society for Gastric Cancer includes so-called linitis plastica and type IV Bormann classification ofthe scirrhous type. 5) This scirrhous type, therefore, is assumed to be the general category for allcarcinoma showing histologically rich interstitia. 5,6)

The etiology of the proliferation of interstitial connective tissue in scirrhous carcinoma isunknown, 1,4, 7) but the connective tissue of the tumor stroma seemingly influences the tumor

growth. 8) We made a hypothesis that carcinoma cell products infiltrating to the interstitia may

induce stromal reaction such as scirrhous type whether the products are specific or non-specificto malignant cells. We us~d an immunofluorescent staining procedure in an attempt at clarifica­

tion. The antiserum we used was presumed to be reactive against products commonly secretedfrom both cancerous and non-cancerous gastric epithelial cells. Substances which relate to scir­rhous carcinoma were discussed with reference to the observation of interactions between suchcellular secretions and the interstitial connective tissue.

Supported by Ministry of Health and Welfare, Imanaga Foundation and Daiwa Health Foundation.#*.,.~ ~.~m8~.~. _,~~.~

Received for publication April 9, 1984

23

24 HID EO KAMEl et of.

MATERIALS AND METHODS

Preparation ofantigen

Gastric mucosa which seemed to be normal and which had been obtained during surgery forduodenal ulcer was used for the preparation of the antigen. The specimen was finely cut with apair of scissors and homogenized in a micro-homogenizer at a ratio of 2 g. of tissue to 5 ml. ofphosphate-buffered saline. After centrifugation at 1800 g for 15 minutes, the supernatant wasused as the antigen. 9) Similar preparations were made from the tissue of gastric carcinoma andjejunal mucosa. One g. of lyophilized extract was reconstituted to 10 ml. in phosphate-bufferedsaline before use as an antigen.

Preparation ofantibody

The antigen preparation from gastric mucosa was then emulsified with Freund's complete ad­juvant. The antigen (0.25 ml) was injected at each of 4 sites in rabbit foot-pads and subcutaneoustissue, and the same antigen emulsion was injected repeatedly at 2-week intervals for 2 months.The rabbit blood was collected 1 week after the booster shot to obtain the antiserum. This anti­serum was absorbed by the human blood cells (AB type), the human liver tissue, and the humanjejunal mucosa, and finally saturated with a 30% ammonium sulfate solution to obtain a -y-globu­lin-rich preparation in PBS. The preparation was dialized in phosphate buffered saline and appliedto 2 X 30 cm. columns containing Sephadex G-25.1!) Final protein concentration was adjustedto 30 mg/ml, and the protein preparation was stored at -80°.

Immunofluorescent technique

The gastric carcinoma tissue and the non-cancerous gastric mucosa extracted peroperativelywere immediately frozen at -20°C. Tissue sections 5 Mm in thickness were prepared by the useof a cryostat and submitted for indirect immunofluorescence staining as follows. Anti-gastricmucosa protein fraction (diluted 10 times in phosphate buffered saline) was placed on the ace­tone-fixed sections. After incubation for 1 hour at 37°C in 100% humidity, the sections wererinsed for 20 minutes in phosphate buffered saline. They were then stained by isothiocyanate­labeled goat immunoglobulin fraction against rabbit 'Y-globulins (available from Hoechst A.G.)by incubation for 1 hour at 37°C. After rinsing for 20 minutes in phosphate buffered saline, theywere studied under a fluorescent microscope (Chiyoda Kogaku, Tokyo).

All specimens obtained from 28 patients were stained by the above antiserum. Of the 28 pa­tients, 4 were cases of gastric ulcer, 10 of non-scirrhous gastric carcinoma, 8 of scirrhous gastriccarcinoma, 3 of early stage gastric carcinoma with ulceration of the epithelium, and the remain­ing 3 of colon carcinoma.

RESULTS

As illustrated in Fig. I, the antiserum to gastric mucosa, before absorption by jejunal mucosaand red blood cell, produced multiple precipitin lines against preparations from gastric carci­noma, gastric mucosa, jejunal mucosa, gastric juice and human serum. Precipitin lines of gastriccarcinoma (in the top well) were contiguous with those of gastric mucosa (in the right upperwell). After absorption by jejunal mucosa, the antiserum produced two precipitin lines each, withgastric mucosa and gastric carcinoma (Fig. 2).

Fig. 3 illustrates a fluorescence-stained specimen of gastric carcinoma cells and non-cancerousgastric epithelial cells. The cytoplasm is well stained in both types of cells.

STROMA OF SCIRRHOUS CARCINOMA

Fig. 1 In the central well, antiserum against gastric mucosa after absorption by human liver tissue wasplaced. The preparation from gastric carcinoma was placed in the top well, that from gastricmucosa in the right upper well, that from jejunal mucosa in the right lower well, that fromgastric juice in the bottom and left lower wells and that from human 0 type serum in the leftupper well.

Fig. 2 In the central well, antiserum against gastric mucosa after absorption by human liver tissue, ABtype red blood cells and jejunal mucosa was placed. The preparation from gastric mucosa wasplaced in the top well, that from jejunal mucosa in the right well, that from gastric carcinoma inthe bottom well and that from gastric juice in the left well.

25

26 HIDEO KAMEl et al.

Fig. 3 Fluorescence-stained tissue imprint of gastric carcinoma (right side) and non-malignant gastricepithelial cells (left side), XIOOO.

Fig. 4 depicts a fluorescence-stained specimen from the gastric mucosal epithelium. The epi­thelial cells are moderately stained; and the lumens of the glands are well stained. Fig. 5 is fluo­rescence-stained non-malignant epithelium showing metaplasia. All of the tested non-malignantepithelial cells and glandular lumens were well stained by this immunoglobulin fraction.

Fig. 6 illustrates a fluorescence-stained specimen of non-scirrhous tubulary adenocarcinoma.Stained substances are dominant in the cytoplasm and in the lumen of the glands.

Fig. 7 represents a hematoxylin-eosin-stained specimen of interstitium-rich carcinoma withdissociated glandular formation. Fig. 8 illustrates the same specimen and almost the same area asFig. 7. The tumor stroma shows a large amount of hyalinized connective tissue. The stained sub­stances are visible in the cytoplasms and to some degree in the stroma as well. Fig. 9 is a fluores­cence-stained specimen from the same material, but the area illustrated represents a rather smallamount of tumor stroma with less hyalinization taken from the lesion adjacent to that shown inFigs. 7 & 8. This is characterized not only by the stained cells but also the lumen of the glands.Fig. 10 illustrates an ulcerated area of early stage gastric carcinoma. Scattered malignant cellsand, seemingly, their products are stained. No such stainability was observed in the benign ulcernor in non-scirrhous carcinoma without ulceration.

As controls, non-scirrhous gastric carcinoma was stained only by fluorescence-labeled goatanti-rabbit 'Y-globulin serum as illustrated in Fig. 11. No stainability was observed in the cyto­plasm or in the lumen of the glands, which contrasted well with Fig. 6. Scirrhous carcinoma(Figs. 8 & 9) was stained by fluorescence-labeled anti-rabbit immunoglobulin after staining withnon-labeled anti-gastric mucosa immunoglobulin fraction which was absorbed by gastric mucosapreparation (Fig. 12), or by fluorescence-labeled anti-gastric mucosa 'Y-globulin after absorption byanti-gastric mucosa 'Y-globulin (Fig. 13). The interstitia were not stained as shown in Fig. 12 & 13.

STROMA OF SCIRRHOUS CARCINOMA

•. ' " ~ '."', '.O ..... . ~';'L • f ',; ',~

\'- "-.-" 1: ,..'1" '.~•." .... ,:>.~. }', -'.' '. r": -\ "... '". • , .. ' -.: , .~.. \. '~

..•. &-Fig. 4 Fluorescence-stained gastric mucosa taken from the non-malignant patients, X250.

Fig. 5 Fluorescence-stained non-malignant gastric mucosa with metaplasia, specimen taken from epi­thelium adjacent to carcinoma lesion, X250.

27

28 HIDEO KAMEl et at.

Fig. 6 Fluorescence-stained non-scirrhous tubular adenocarcinoma, X500.

Fig. 7 Markedly dissociated adenocarcinoma with rich interstitium, H.E. stain, X250.

STROMA OF SCIRRHOUS CARCINOMA

Fig. 8 Fluorescence staining of a lesion similar to that shown in Fig. 7, X250.

Fig. 9 Fluorescence staining of the lesion adjacent to the area of Figs. 7 & 8. The glands are moder­ately dissociated with rather small amount of interstitium, X250.

29

30 HIDEO KAMEl et ai.

Fig. 10 Superficial ulcer of early gastric carcinoma. Scattered cancer cells and their products in thestroma were visible. XSOO.

Fig. 11 Non-scirrhous gastric carcinoma stained directly by fluorescence-labeled goat anti-rabbit -y_globulin serum, XSOO.

STROMA OF SCIRRHOUS CARCINOMA

Fig. 12 Scirrhous type gastric carcinoma stained by immuno-fluorescence-labeled anti-rabbit immuno­globulin after staining with non-labeled antigastric mucosa immunoglobulin fraction whichwas absorbed by gastric mucosa preparation, X250.

Fig. 13 Stained by immunofluorescence-labeled anti-gastric mucosa immunoglobulin fraction afterthe absorption by non-labeled anti-gastric mucosa serum, X250 (similar area with Fig. 9)

31

32 HIDEO KAMEl et ai.

As shown in table 1, metastatic lesions of gastric carcinoma in the lymph nodes and in bothmalignant and non-malignant colon epithelial cells were positively stained with this antiserum.Other organs tested by this antiserum were negative for staining. Gastric mucosa of mice and ratswere also not stainable.

Table 2 shows the result of stainablility of the cells, the lumen of glands and the interstitia in28 cases by the use of this immunoglobulin fraction. Gastric epithelial cells were mostly stainedpositive whether they were non-malignant or malignant. Colon epithelium of normal and malig­nancy were also more or less stainable. Glandular lumens were well stained except for the scir­rhous type, while interstitia were stainable in the ulcerated area, where dissociated glandulararrangement was observed even at early stages, and in the scirrhous type. In both cases of med­ullary gastric carcinoma without glandular arrangement, well-stained substances were observedin the cells but not in the interstitia. In table 2, non-scirrhous gastric carcinoma included 2 casesof early carcinoma without formation of ulcer whose interstitia were not stained by this immuno­globulin fraction.

Table 1. Result of the immunofluorescent staining of various human organs and gastric mucosa ofanimals by the use' of antihuman gastric mucosa 'l'-globulin fraction.

Organ Number of cases Number showing grade of staining*± +

esophagus 3 3 0 0

small intestine 3 3 0 0large intestine 4 2 2 0

metastasis of gastric carcinoma 4 0 0 4to lymph node

metastasis of colon carcinoma3 0 3 0to lymph node

lymph nodes 4 4 0 0

spleen 3 3 0 0

liver 3 3 0 0

pancreas 2 2 0 0

skin 2 2 0 0

gastric mucosa of rat 3 3 0 0

gastric mucosa of mouse 3 3 0 0

* -: negative for stain ±: positive but trace for stain +: definitly positive for stain

Tissue from organs was obtained during surgery. Gastric mucosa of rat and mouse was stained shortlyafter excizing under anesthesia by ether.

STROMA OF SCIRRHOUS CARCINOMA

Table 2. Grade of stain in various alimental lesions by the use of anti-human gastric mucosa-y-globulin fraction.

Case Number of cases Cell Glandular Stromalumen

+ ++ + ++ + ++

gastric ulcer (epithelim taken4 0 0 4 0 0 4 4 0 0apart from ulcer)

gastric ulcer (lesion of ulcer4without epithelization)

non-scirrhous gastric carcinoma 8 0 3 5 0 2 6 7 0

medullary gastric carcinoma2 0 2 0 0without glandular formation

scirrhous gastric carcinoma 8 0 2 6 3 5 0 0 4 4

early gastric carcinoma with3 0 2 0 3 0ulceration

colon carcinoma 3 2 0 2 0 3 0 0

-: negative for stain +: moderately positive ++: strongly posi tive

Figure indicates number of cases.

DISCUSSION

33

As observed in Figs. I & 2, this immunoblobulin fraction, after its absorption by the humanliver tissue, produced multiple precipitin lines with the preparations from gastric carcinoma,jejunal mucosa and gastric juice. Since precipitin lines are only observed with the preparationsfrom gastric carcinoma, gastric mucosa and gastric juice after absorption by jejunal mucosa andare contiguous with each other, they are seemingly specific to cell components and products ofthe gastric epithelium regardless of whether these are normal or malignant. However, other anti­bodies which are not detectable by the precipitin test but which are sensitive to immunofluores­cent staining may be presented in this fraction. This immunoglobulin fraction gave positive im­muno-fluorescence staining to cells and their products of gastric origin but not to those of otherorgans nor to heterologous organs, the only exception being colon epithelium which produced amore or less cross reaction with this fraction. These results strongly suggest that this fraction isspecific to gastrointestinal natures. It is well known that malignant cells and their products keeporgan-specific antigenicities although these may decrease during malignant transformation,u)Recently, many monoclonal antibodies have been detected against normal and non-malignantcells, and most of them are against organ or tissue-specific antigens or stage-specific embryonicantigens. 12,13 ,14,15) Even though these monoclonal antibodies produce positive staining in thestroma, the reacting antigens can not be the inducers of stromal reaction. The immuno-globulinfraction used in this study was not monospecific; however, the object of observation is whetheror not infiltration of cell products induces the proliferation of interstitial connective tissue ingastric carcinoma.

It is characteristic that fluorescence-stained location in scirrhous carcinoma was specific com­pared to other types of gastric carcinoma and benign epithelium. In scirrhous carcinoma, therewas fluorescent staining in the stroma and slight staining in the lumen of the glands. This sug­gested that infiltration of the interstitium by glandular antigens occurred. This may be explainedby the fact that dissociated glands no longer communicate with the viscus cavity.4,8) In scirrhousgastric carcinoma, signet cells are frequently observed at the mucosal layer. However, the type ofcell at the mucosal layer is not always the same as that observed in the submucosal layer. This

34 HIDEO KAMEl et al.

seemingly explains why the microscopic cell type is not specific to scirrhous carcinoma. In earlygastric carcinoma, on the other hand, signet ring cells are frequently observed in case of the ulcer­forming type. S,16) There is speculation that early gastric carcinoma of ulcer-forming type maydevelop into scirrhous gastric carcinoma. 16) The results showing that immunofluorescence stain­ing was well visualized in the stroma of scirrhous carcinoma and in the ulcerating area of earlygastric carcinoma strongly suggested that these infiltrating substances induce scirrhous typegastric carcinoma.

REFERENCES

1) Anderson, W.A.D. and Kissane, J.M., Pathology, 7th ed., C.V Mosby Comp., St. Louis pp. 1306-1309,1977.

2) Flatau, E. et al., Linitis plastica inmtrating the entire gut. Am. J. Gastroenterol., 77,559-561, 1982.3) Robbins, S.L. and Cotran, R.S., Pathologic Basis of Disease, 2nd ed., W.B. Saunders Comp., Philadelphia,

pp. 946-951,1979.4) Rosai, J., Ackerman's Surgical Pathology, 6th ed., C.V. Mosby Comp., St. Louis, pp. 434-443,1972.5) Japanese Research Society for Gastric Cancer, The General Rules for the Gastric Cancer Study in Surgery

and Pathology, 10th ed., Kanehara Shuppan, Tokyo, 1979.6) Kamei, H. et al., Mitotic index in gastric carcinoma of the scirrhous type. Jpn. J. Surg., 11,337-340,

1981.7) Fernet, P., Azar, H.A. and Stout, A.P., Intramural spread of linitis plastica along the alimentary tract.

Gastroenterology, 48,419-424,1965.8) Kamei, H., Relationship between tumor growth and the environmental connective tissue. Nagoya J. Med.

Sci., 27, 142-166, 1964.9) Tosner, J. and Fixa, B., Non-specific leucocyte migration inhibition of gynaecological carcinoma patients,

Clin. Exp. Immunol., 38,356-360,1979.10) Kamei, H. and Hoshikawa, S., Identification of lymphoid cells by the use of immunofluorescent staining

with antiserum against cultured lymphoblastoid cells. Jpn. J. Surg., 2,73-76, 1972.11) Kawasaki, H. and Kimoto, E., Mucosal glycoproteins in carcinoma cells ofgastrointestinal tract, as detected

by immunofluorescence technique. Acta Path. Jpn., 24;481-494, 1974.12) Bale, W.F., Contreras, M.A. and Grady, E.D., Factors influencing localization of labeled antibodies in

tumors. Cancer Res., 40, 2965 -2972, 1980.13) Goldenberg, D.M., Sharkey, R.M. and Primus, F.J., Carcinoembryonic antigen in histology: Immuno~

peroxidase staining of conventional tissue sections. J. Nat!. Cancer Inst., 57, 11-22, 1976.14) Gold, P. and Shuster, J., Historical development and potential uses of tumor antigens as markers of human

cancer growth.. Cancer Res., 40,2973-2976,1980.IS) Ballou, B. et al., Tumor radioimmunolocation: Differential antibody retention by antigenic normal tissue

and tumor. J. Immunol., 132,2111-2116, 1984.16) Kino, I., Histological typing and prognosis of the carcinoma of the alimentary tract. Cancer and Chemo·

therapy, 4, 1203-1209, 1977.